There is no single, clean number for how many deaths are caused by driving high on marijuana each year, and the reason matters. Unlike alcohol, where blood levels reliably indicate impairment, THC testing in crash victims only proves recent use, not that the driver was impaired at the moment of the crash. That said, the data we do have paints a concerning picture: by 2018, 21.5% of all U.S. motor vehicle crash fatalities involved a driver who tested positive for cannabis, more than double the 9% recorded in 2000.
With roughly 40,000 to 43,000 traffic deaths per year in the U.S., that percentage translates to somewhere around 8,000 to 9,000 fatalities annually involving a cannabis-positive driver. But “involving” is not the same as “caused by,” and untangling that distinction is central to understanding the real risk.
Why the Numbers Are Hard to Pin Down
Alcohol impairment is straightforward to measure. A blood alcohol concentration of 0.08% means a person is legally drunk, and decades of research confirm that level degrades driving ability in predictable ways. THC doesn’t work like that. It’s fat-soluble, meaning it can linger in blood and urine for days or even weeks after the impairing effects have worn off. A driver who tests positive for THC after a fatal crash may have smoked that morning, the night before, or a week earlier.
Research has found that drivers with blood THC concentrations above 3 nanograms per milliliter are more likely to show observable impairment, but the range is enormous. In one study of apprehended drivers, THC levels in impaired individuals ranged from 0.3 to over 45 nanograms per milliliter. There’s no threshold that works the way 0.08% does for alcohol, and tolerance plays a large role: regular users show less impairment at the same blood levels than occasional users do.
This makes it genuinely difficult for researchers, and for law enforcement, to say with certainty how many crashes cannabis directly caused versus how many simply involved a driver who had used cannabis at some point recently.
What Cannabis Actually Does to Driving
The impairment is real, even if the measurement is messy. Cannabis primarily affects complex decision-making rather than basic motor skills. In driving simulator studies, THC significantly increased lane weaving within the first 40 to 100 minutes after use. Drivers drifted out of their lane more often and showed slower reaction times, particularly in situations requiring split-second decisions like braking for a red light.
Interestingly, simple tasks were relatively unaffected. Drivers could handle straightforward processing, like identifying a single object on a screen, without measurable impairment. But when tasks got more complex, requiring divided attention or rapid visual scanning, performance dropped significantly. This matters because real-world driving constantly demands that kind of multitasking: checking mirrors, reading traffic flow, spotting a pedestrian stepping off a curb while also monitoring speed.
The peak impairment window appears to be roughly the first two to four hours after smoking or vaping, though edibles can shift that window later and make it less predictable.
Cannabis Alone vs. Cannabis With Alcohol
One of the most important findings in crash research is how dramatically risk escalates when cannabis and alcohol are combined. A case-control study comparing crash-involved drivers to the general driving population found that cannabis alone, without alcohol, increased the odds of a fatal crash by about 1.5 times. That’s a modest increase, roughly comparable to driving with a passenger who distracts you.
Alcohol alone increased fatal crash odds by about 16 times. But drivers who tested positive for both alcohol and cannabis had 25 times the odds of a fatal crash compared to sober drivers. The combination is far worse than either substance on its own, and it’s not rare: by 2018, about 10.3% of all traffic fatalities involved drivers positive for both cannabis and alcohol, up from 4.8% in 2000.
Who Is Most Likely to Drive High
Young adults between 21 and 25 report the highest rates of driving under the influence of marijuana, at 12.4%. Drivers aged 16 to 20 come in second at 9.2%, a particularly concerning figure because teen drivers already face elevated crash risk due to inexperience. The rate drops steadily with age, falling to less than 1% among drivers 65 and older.
Men are roughly twice as likely as women to drive after using marijuana (6.2% versus 3.2%). These patterns mirror alcohol-impaired driving demographics closely, suggesting that the same risk tolerance and decision-making patterns drive both behaviors.
How Legalization Has Affected Crash Rates
Research from the Insurance Institute for Highway Safety found that states legalizing recreational marijuana saw a 4.1% increase in fatal crash rates when combining the effects of legalization and the start of retail sales. Injury crashes rose by 5.8%. The increase appeared to happen around the time of legalization itself rather than when retail shops opened, suggesting that the cultural shift in attitudes toward use mattered more than access to dispensaries.
The effects weren’t uniform across states. Fatal crash rate changes ranged from a 10% decrease in some states to a 4% increase in others. Local factors like enforcement intensity, road design, and existing rates of cannabis use all play a role. Still, the overall trend points to a small but real increase in traffic deaths associated with legalization.
Why Enforcement Lags Behind
Part of the challenge in reducing cannabis-impaired driving deaths is that there’s no equivalent of a breathalyzer for marijuana. Roadside oral fluid tests can detect recent THC use with about 79% sensitivity and 98% specificity, meaning they catch most users but miss roughly one in five. More importantly, a positive test doesn’t tell an officer whether the driver is actually impaired right now.
Many states rely on drug recognition experts, officers trained to identify signs of impairment through physical tests and observations. But this is subjective and resource-intensive compared to the simple, reliable breath test used for alcohol. Several states have set legal limits for THC blood concentration (commonly 5 nanograms per milliliter), but these thresholds don’t map neatly onto impairment the way alcohol limits do.
NHTSA’s National Roadside Survey found that the percentage of weekend nighttime drivers testing positive for marijuana rose from 8.6% in 2007 to 12.6% in the 2013-2014 survey, a period that predates most state legalization. During a study period covering October through December 2020, 56% of drivers involved in serious injury and fatal crashes tested positive for at least one drug at participating trauma centers, though that figure includes all drugs, not just cannabis.
The Bottom Line on the Numbers
Cannabis is present in a growing share of fatal crashes, now appearing in roughly one in five traffic deaths in the U.S. The best available evidence suggests cannabis alone roughly doubles crash risk, a meaningful increase but far smaller than alcohol’s effect. The real danger multiplier comes from combining the two. If you’re trying to put a number on it, somewhere between 2,000 and 4,000 annual U.S. traffic deaths likely involve cannabis as a contributing factor rather than just a bystander in the toxicology report, though honest researchers will tell you that range is an educated estimate, not a precise count.